Research On Individual Moving Feature Recognition Based On Participatory Sensing

Sensing technology and mobile computing enhance breadth and depth of peoplecollecting, analyzing, and making use of the data in an unprecedented way. LocationBased Services (LBS), which includes such things as intelligent transportation, medicalmonitoring of the special groups, aid navigation in complex environment, Individualmobile feature recognition becomes one of the key factors in determining the quality ofLBS. Due to existing research results are limited by mobile devices’ energy and serviceextensibility, they can’t satisfy individual obvious differences and mobile environmentalservices requirements when environment is dynamic. We conduct the research on twoimportant mobile future including movement behavior recognition and neighbor discoverybased on participatory sensing of mobile users.Firstly, we propose a kind of statistical sensing method for low-power movementbehavior recognition. We analyze the temporal statistical characterization based on thesensory samples for both the moving behavior and the stationary one. We use Bayes’theorem to distinguish between these moving and stationary behaviors. Moving behaviors(walking, running, cycling or motoring), can be identified in terms of combining themobility with their statistical characterization amongst various behaviors.Secondly, we propose the environmental characteristics participatory extractionmethod for moving individual neighbor discovery. We use lightweight sensors to collectdata. We use SVM (Support Vector Machine), Tanimoto Coefficient and Manhattandistance to calculate three kinds of fingerprint similarity respectively, and then theprincipal component analysis is used to reduce data dimension in order to obtain neighborsimilarity rank. It satisfies neighbor discovery real-time, universality and accuracy on thepremise of guarantee not to reveal the user privacy.Lastly, we implement different types of mobile phones to collect real sensor data andthen use Matlab to design complete simulation experiment, which verify the proposedmovement behavior recognition and moving individual neighbor discovery results.